Is Your Website Ready for AI Search and AI Agents?
AI-powered search engines and browser agents interact with websites differently from traditional visitors. This free AI Readiness Checker examines whether your website exposes accessible content, provides useful structured information and supports the discovery and interaction mechanisms AI systems look for. Enter a URL to get a diagnostic score out of 100, the issues behind it and what to do about each one.
No signup required
Free diagnostic report
No payment required
Actionable recommendations
What gets checked
Seven areas, one AI Readiness Index
Each area carries a fixed share of the score. Every check inside it comes back as a pass, a warning or a failure, with the evidence that produced it.
AI crawler access
An AI Crawler Accessibility Checker for your robots.txt: whether OAI-SearchBot, ChatGPT-User, PerplexityBot, Googlebot and other AI crawlers may fetch the page, plus noindex rules and firewall blocks.
Content AI can read
How much text is in the HTML before any JavaScript runs - most AI crawlers do not run it - plus the title, meta description, one H1, heading structure and page language.
Structured data and entities
Valid JSON-LD, an Organization and WebSite entity that says who is behind the site, page-type schema such as Article, Product or FAQPage, and Open Graph tags.
AI discovery files
An llms.txt file that summarises your site for language models, an XML sitemap, the sitemap listed in robots.txt, and a canonical URL.
GEO and AEO signals
Our GEO Readiness Checker and AEO Score Checker look for question-style headings, an FAQ, lists and tables answer engines can lift, visible dates and clear authorship.
AI agent readiness
An Agentic Web Readiness Checker for what AI agents need to browse and act: HTTPS, a fast response, semantic landmarks, labelled fields, named buttons and image alt text.
Agent interfaces
Machine-readable interfaces an AI agent looks for: an API catalogue, OAuth discovery, an MCP server card and agent skill files.
These checks measure technical and content signals that AI crawlers, answer engines and agents are known to use. They cannot measure whether an AI system will actually quote your page: crawler accessibility is one part of a broader discoverability strategy, not the whole of it.
Why it matters
Three things an AI system does with your website
AI search engines, answer engines and browser agents each use a site differently from a human visitor, and differently from one another. These are the three jobs the checker measures, and what each one needs from your pages.
Access
It has to be able to fetch the page at all
An AI crawler asks your server for the HTML and reads what comes back. It obeys robots.txt, a firewall rule can stop it before it ever reaches your content, and - the part that surprises people most - the majority of AI crawlers do not run JavaScript. If your text only appears once a framework has hydrated, the crawler sees an empty shell.
robots.txt is read per crawler, and the two groups are not the same decision: blocking GPTBot keeps you out of training data, blocking OAI-SearchBot keeps you out of answers.
Text the server already sent is the only text an AI crawler is guaranteed to see.
A CDN or bot-fight rule can refuse AI crawlers even when robots.txt welcomes them, so we request your page as one and compare the answers.
It has to work out what the page says, and who is saying it
An answer engine quotes a passage. A generative engine names a source. Both need content they can lift in one piece and an identity they can attribute it to - which is the whole of what GEO and AEO describe: question-shaped headings and direct answers for the first, structured data, dates and clear authorship for the second.
A heading that asks the question and a paragraph that answers it directly underneath is the shape an answer engine can quote.
JSON-LD states the facts a machine would otherwise have to infer: who you are, what kind of page this is, when it last changed.
A named publisher and a visible date are a large part of how a generative system decides a source is worth citing at all.
It has to be able to use the page, not only read it
Browser agents click links, fill in forms and finish tasks on someone's behalf. They work from the accessibility tree, so an unlabelled field or an icon button with no name is simply not there as far as they are concerned. A newer layer goes further still: machine-readable documents that describe an interface directly instead of making an agent drive the screen.
Landmarks, labelled fields and named buttons are what an agent navigates by - the same things a screen reader needs, which is why accessibility work pays twice.
HTTPS and a quick first response, because an agent working through a multi-step task gives up on a slow page.
Emerging conventions - an API catalogue, OAuth discovery, an MCP server card, agent skill files - which we report on without marking anyone down for being early.
Nothing below is a badge or a slogan. Each statement describes what the scanner actually does, and the pages linked at the end say the same thing in more detail.
48
Checks per page
Each one reports the evidence behind it.
7
Weighted areas
Every check belongs to exactly one.
0
Reports stored
The free single-page check keeps no copy.
Who builds it
Rudra Analyzer is developed and run by Rudra Techno. The same team answers the support address below.
What the scanner tests
Plain HTTP requests: your page, your page again with an AI crawler's user agent, robots.txt, llms.txt, your sitemap and a short fixed list of well-known agent files. Nothing is rendered, no JavaScript is run, no form is submitted and no credentials are ever sent.
What is collected during a scan
The address you enter and the responses your server gives back. We do not sign in to anything, and our analytics never receives the address you submitted.
Whether reports are stored
The free single-page check stores nothing: it runs inside your request, the result is returned to your browser, and there is no saved report, no shareable link and no history. A signed-in whole-site check is deleted automatically 24 hours after its last activity.
No proof of ownership needed
The check only reads pages that already load publicly, so it needs no verification that the site is yours — and it cannot see anything an ordinary visitor could not.
What we will not claim
Every finding shows its evidence — the blocked bot names, the word count, the status code — so you can verify it yourself. We publish no customer counts, ratings, logos or before-and-after numbers, because we have no honest way to measure them.
Scanning a URL is not access to your code
The free check reads public pages over HTTP. It never touches a code repository and never asks you to connect one.
The optional paid fix is the only part that involves your code, and it starts only after you select the pages and issues, review the scope, confirm the price and explicitly authorize access to one repository. We read only that repository. Changes arrive on a separate branch for you to review; merging and deploying require your own separate authorization.
We never ask for a GitHub password, a personal access token, a private key or any repository secret.
The report is free and complete — no finding is held back behind a payment. Having the work done for you is the only paid, optional part.
1
Enter your website URL
One page: your home page, or a page you would want quoted in an AI answer. No account and no card.
2
Run the free AI readiness scan
We make a handful of read-only requests and score the answers. It takes about 10 seconds.
3
Review the score and prioritized findings
A score out of 100, a score for each of the seven areas, and every finding ordered by priority with the evidence, why it matters and how to fix it.
4
Fix issues yourself, or request implementation
Work through the report with your own developer at no cost, or have our team implement the eligible findings in your codebase for a price you confirm first.
These checks are free and so are the guides. If you'd prefer a person to make the changes, our team offers paid help.
Review the reportrun the free AI Readiness check and read the findings.
2
Choose the scopepick the pages and the findings you want fixed.
3
Confirm the pricethe page count, the tier and the total are shown before you pay.
4
Pay oncea single payment through Stripe or Razorpay. Nothing recurring.
5
Authorize accessinstall our GitHub App on that one repository, or invite a collaborator.
6
A branch of its ownwe record your default branch and its commit, then branch from it.
7
Implement the fixesonly the findings in the agreed scope, only in files that need changing.
8
Check the workevery changed file is read back and re-checked before you see it.
9
A change summarywhat was fixed, what could not be verified, and what still needs a person.
10
You reviewread the diff or the pull request, and merge when you're satisfied.
11
Merge and deploy (optional)authorize each separately, and your own CI and branch rules still apply.
What it costs
One payment, priced by how many distinct pages are in the agreed scope - not by how many findings there are.
Loading prices…
This is a one-time payment. We do not store a payment method, nothing renews, and no amount is ever taken without you starting that payment.
A page is a distinct HTML page. Images, scripts and stylesheets are not counted, and the same page reached by several addresses is counted once.
These tiers buy a bounded amount of work. If your site needs more than the plan covers, we stop and tell you - we will never widen the scope or charge again without you agreeing to it first.
How we get access to your code
GitHub App (preferred)
You install our app on the one repository you want fixed, with only the permissions needed to create a branch, commit the approved changes and open a pull request. Access tokens are short-lived and never stored.
Manual collaborator invitation
Invite our GitHub account to the repository with Write access, and remove it again when the work is done.
Verify the GitHub account before granting access to a private repository. The support email address is not a GitHub username - do not invite an account you have not checked.
We never ask for a personal access token, a password, a private key or any repository secret. We do not read unrelated repositories, and files that can hold credentials are never edited or committed.
The scan and the report are free. The fix service is optional.
Answers for you
AI readiness, answered
What the checker measures, what the score does and does not mean, and where the free report ends and the optional paid service begins.
What is an AI Readiness Checker?
It is a diagnostic tool that tests how well a website works for AI search engines, answer engines and AI agents rather than for human visitors. Ours returns the Rudra Analyzer AI Readiness Index, a score from 0 to 100 combining seven areas: AI crawler access, content AI can read, structured data and entities, AI discovery files, GEO and AEO signals, AI agent readiness, and machine-readable agent interfaces.
Is the AI readiness checker free?
Yes. It needs no account. Checks are limited to a few per hour for each network so everyone gets a turn.
Do I need to register?
No. The single-page check needs no account, no email address and no payment. An account is only needed for the whole-site check, which scores every page in your sitemap at once.
Which AI crawlers are checked?
Two groups, read from your robots.txt. AI search and answer crawlers: OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, PerplexityBot, Perplexity-User, DuckAssistBot, Applebot, Googlebot and Bingbot. AI training crawlers: GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, CCBot, meta-externalagent, Bytespider and Amazonbot. The page is also requested once with an AI crawler's user agent, which catches a firewall or CDN that turns bots away even when robots.txt allows them.
What is a good Website AI Readiness Score?
85 or more means ready for AI, 65–84 almost ready, 40–64 partly ready and below 40 not ready yet. Fix the failed checks first: they carry the most weight.
Does a high AI readiness score guarantee ChatGPT citations?
No, and no tool can promise that. The checker measures signals you control: whether crawlers are allowed in, whether your content is in the HTML, whether it is structured and quotable. Which pages an AI assistant actually cites depends on each provider's own retrieval and ranking, which is not published. Crawler accessibility is a prerequisite, not a guarantee.
What is the difference between GEO and AEO?
Answer engine optimization (AEO) is about being the answer to a specific question: clear question-style headings, a direct answer near the top, FAQs, and lists and tables that can be lifted whole. Generative engine optimization (GEO) is broader — being a source a generative system draws on and names, which also depends on structured data, a clear publisher identity, dates and authorship. They overlap heavily, and this checker scores both in one area.
Does llms.txt improve Google rankings?
No. Google has not said it uses llms.txt, and no search engine has confirmed it as a ranking signal. It is a proposed convention for giving language models a short, readable map of your site. We check for it because it is cheap to publish and some AI tools do read it — not because it is required. Treat it as optional, and never let it displace a working XML sitemap.
Should I block GPTBot and other AI training crawlers?
That is a business choice. Blocking training crawlers only lowers the score slightly. Blocking AI search crawlers such as OAI-SearchBot, ChatGPT-User or PerplexityBot is different: it keeps your pages out of AI answers, so it counts as a failure.
How can I improve my website's AI readiness?
Work down the report by priority. The biggest gains are usually: let AI search crawlers through robots.txt and your firewall; make sure the page's text is in the HTML the server sends, not only after JavaScript runs; add valid JSON-LD naming your organization and the page type; publish an XML sitemap and a canonical URL; and give pages question-style headings, visible dates and clear authorship. The machine-readable interface checks score low for almost every site today — they are emerging standards, so treat them as an opportunity rather than an emergency.
Can I scan individual pages?
Yes. The free check scores one address at a time, so you can test your home page, a pricing page or a single article. With a free account you can run a whole-site check instead, which follows your sitemap and reports findings across every page it scores, up to your plan's page limit.
How can I fix failed checks?
Open the finding in the report. Each one shows what we found — the blocked bot names, the word count, the status code — why it matters and the change to make. You can also copy a ready-made prompt from the result into a coding assistant that can edit your site. If you would rather not do it yourself, the optional paid service implements eligible findings for you.
Is my website data stored?
The free single-page check stores nothing. It runs inside the request, returns the result to your browser and keeps no copy — there is no saved report and no shareable link. A signed-in whole-site check is deleted automatically 24 hours after its last activity. The privacy policy covers everything else we hold, such as account details.
Do you require GitHub access for the free scan?
No. The free check only makes public HTTP requests to your website; it never touches a repository and never asks you to connect one. Repository access is only ever involved in the optional paid fix, and only after you authorize it yourself for one named repository. We never ask for a password or a personal access token.
Can Rudra Analyzer implement the recommended fixes?
Yes, as a separate paid service. You select the pages and issues, review the scope and confirm the price before paying, then authorize access to one repository. We work on a branch of its own and send you a summary of what changed and what still needs a person. You review the pull request before merging, and merging and deployment are always your own separate decision.
How is this different from a website audit or an SEO audit?
A website audit checks speed, SEO, mobile, accessibility, security and links; an SEO audit focuses on search engines. The AI readiness check focuses on AI crawlers, answer engines and agents. Run all three for the full picture - the result page links to both.
Keep reading
Guides and tools for the next step
The findings above each have a guide behind them, and the other free checks look at the things this one deliberately does not.